I think that in the long run (maybe not too long) we will have to change our stateful operators (windows, basically) to use managed memory and spill to disk. (Think jobs that have sliding windows over days or weeks) Then then the internal operators will take care of copying anyways. The problem Gyula mentioned we cannot tackle other than by defining how user code must behave.
On Wed, May 20, 2015 at 3:30 PM, Stephan Ewen <se...@apache.org> wrote: > It does not mean we have to behave the same way, it is just an indication > that well-defined behavior can allow you to mess things up. > > The question is now what is the default mode: > - Failsafe/Heavy (always copy) > - Performance/Lightweight (do not copy) > > > On Wed, May 20, 2015 at 3:29 PM, Stephan Ewen <se...@apache.org> wrote: > >> This is something that we can clearly define as "should not be done". >> Systems do that. >> I think if you repeatedly emit (or mutate) the same object for example in >> Spark, you get an RDD with completely messed up contents. >> >> On Wed, May 20, 2015 at 3:27 PM, Gyula Fóra <gyf...@apache.org> wrote: >> >>> If the preceding operator is emitting a mutated object, or does something >>> with the output object afterwards then its a problem. >>> >>> Emitting the same object is a special case of this. >>> >>> On Wed, May 20, 2015 at 3:09 PM, Stephan Ewen <se...@apache.org> wrote: >>> >>> > The case you are making is if a preceding operator in a chain is >>> repeatedly >>> > emitting the same object, and the succeeding operator is gathering the >>> > objects, then it is a problem >>> > >>> > Or are there cases where the system itself repeatedly emits the same >>> > objects? >>> > >>> > On Wed, May 20, 2015 at 3:07 PM, Gyula Fóra <gyf...@apache.org> wrote: >>> > >>> > > We are designing a system for stateful stream computations, assuming >>> long >>> > > standing operators that gather and store data as the stream evolves >>> > (unlike >>> > > in the dataset api). Many programs, like windowing, sampling etc hold >>> the >>> > > state in the form of past data. And without careful understanding of >>> the >>> > > runtime these programs will break or have unnecessary copies. >>> > > >>> > > This is why I think immutability should be the default so we can have >>> a >>> > > clear dataflow model with immutable streams. >>> > > >>> > > I see absolutely no reason why we cant have the non-copy version as an >>> > > optional setting for the users. >>> > > >>> > > >>> > > On Wed, May 20, 2015 at 2:21 PM, Paris Carbone <par...@kth.se> wrote: >>> > > >>> > > > @stephan I see your point. If we assume that operators do not hold >>> > > > references in their state to any transmitted records it works fine. >>> We >>> > > > therefore need to make this clear to the users. I need to check if >>> that >>> > > > would break semantics in SAMOA or other integrations as well that >>> > assume >>> > > > immutability. For example in SAMOA there are often local metric >>> objects >>> > > > that are being constantly mutated and simply forwarded periodically >>> to >>> > > > other (possibly chained) operators that need to evaluate them. >>> > > > >>> > > > ________________________________________ >>> > > > From: Gyula Fóra <gyf...@apache.org> >>> > > > Sent: Wednesday, May 20, 2015 2:06 PM >>> > > > To: dev@flink.apache.org >>> > > > Subject: Re: [DISCUSS] Re-add record copy to chained operator calls >>> > > > >>> > > > "Copy before putting it into a window buffer and any other group >>> > buffer." >>> > > > >>> > > > Exactly my point. Any stateful operator should be able to implement >>> > > > something like this without having to worry about copying the object >>> > (and >>> > > > at this point the user would need to know whether it comes from the >>> > > network >>> > > > to avoid unnecessary copies), so I don't agree with leaving the copy >>> > off. >>> > > > >>> > > > The user can of course specify that the operator is mutable if he >>> wants >>> > > > (and he is worried about the performance), But I still think the >>> > default >>> > > > behaviour should be immutable. >>> > > > We cannot force users to not hold object references and also it is a >>> > > quite >>> > > > unnatural way of programming in a language like java. >>> > > > >>> > > > >>> > > > On Wed, May 20, 2015 at 1:39 PM, Stephan Ewen <se...@apache.org> >>> > wrote: >>> > > > >>> > > > > I am curious why the copying is actually needed. >>> > > > > >>> > > > > In the batch API, we chain and do not copy and it is rather >>> > > predictable. >>> > > > > >>> > > > > The cornerpoints of that design is to follow these rules: >>> > > > > >>> > > > > 1) Objects read from the network or any buffer are always new >>> > objects. >>> > > > > That comes naturally when they are deserialized as part of that >>> (all >>> > > > > buffers store serialized) >>> > > > > >>> > > > > 2) After a function returned a record (or gives one to the >>> > collector), >>> > > > it >>> > > > > if given to the chain of chained operators, but after it is >>> through >>> > the >>> > > > > chain, no one else holds a reference to that object. >>> > > > > For that, it is crucial that objects are not stored by >>> > reference, >>> > > > but >>> > > > > either stored serialized, or a copy is stored. >>> > > > > >>> > > > > This is quite solid in the batch API. How about we follow the same >>> > > > paradigm >>> > > > > in the streaming API. We would need to adjust the following: >>> > > > > >>> > > > > 1) Do not copy between operators (I think this is the case right >>> now) >>> > > > > >>> > > > > 2) Copy before putting it into a window buffer and any other group >>> > > > buffer. >>> > > > > >>> > > > > >>> > > > > >>> > > > > >>> > > > > >>> > > > > >>> > > > > >>> > > > > >>> > > > > On Wed, May 20, 2015 at 1:22 PM, Aljoscha Krettek < >>> > aljos...@apache.org >>> > > > >>> > > > > wrote: >>> > > > > >>> > > > > > Yes, in fact I anticipated this. There is one central place >>> where >>> > we >>> > > > > > can insert a copy step, in OperatorCollector in OutputHandler. >>> > > > > > >>> > > > > > On Wed, May 20, 2015 at 11:17 AM, Paris Carbone <par...@kth.se> >>> > > wrote: >>> > > > > > > I guess it was not intended ^^. >>> > > > > > > >>> > > > > > > Chaining should be transparent and not break the >>> correct/expected >>> > > > > > behaviour. >>> > > > > > > >>> > > > > > > >>> > > > > > > Paris? >>> > > > > > > >>> > > > > > > On 20 May 2015, at 11:02, Márton Balassi <mbala...@apache.org >>> > >>> > > > wrote: >>> > > > > > > >>> > > > > > > +1 for copying. >>> > > > > > > On May 20, 2015 10:50 AM, "Gyula Fóra" <gyf...@apache.org> >>> > wrote: >>> > > > > > > >>> > > > > > > Hey, >>> > > > > > > >>> > > > > > > The latest streaming operator rework removed the copying of >>> the >>> > > > outputs >>> > > > > > > before passing them to chained operators. This is a major >>> break >>> > for >>> > > > the >>> > > > > > > previous operator semantics which guaranteed immutability. >>> > > > > > > >>> > > > > > > I think this change leads to very indeterministic program >>> > behaviour >>> > > > > from >>> > > > > > > the user's perspective as only non-chained outputs/inputs >>> will be >>> > > > > > mutable. >>> > > > > > > If we allow this to happen, users will start disabling >>> chaining >>> > to >>> > > > get >>> > > > > > > immutability which defeats the purpose. (chaining should not >>> > affect >>> > > > > > program >>> > > > > > > behaviour just increase performance) >>> > > > > > > >>> > > > > > > In my opinion the default setting for each operator should be >>> > > > > > immutability >>> > > > > > > and the user could override this manually if he/she wants. >>> > > > > > > >>> > > > > > > What do you think? >>> > > > > > > >>> > > > > > > Regards, >>> > > > > > > Gyula >>> > > > > > > >>> > > > > > > >>> > > > > > >>> > > > > >>> > > > >>> > > >>> > >>> >> >>
